Bid invitation review data processing method and device, storage medium and computer equipment

Abstract

The invention discloses a bid invitation review data processing method and device, a storage medium and computer equipment. The method comprises the following steps: in response to a bid invitation signal of a target bid invitation project, obtaining to-be-reviewed data of a bidder, and determining a plurality of review indexes of the target bid invitation project; based on each review index, classifying the to-be-reviewed data to obtain to-be-reviewed data in different categories; inputting the to-be-reviewed data under each category into a preset scoring model for scoring to obtain a prediction score under each review index; determining a review terminal corresponding to each review index, and sending the to-be-reviewed data corresponding to each review index to the corresponding review terminal, so as to determine a terminal score under each review index based on the corresponding to-be-reviewed data through each review terminal; and receiving the terminal score of each review index, and determining a review result of the bidding party based on the predicted score and the terminal score under each review index.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular, to a method, device, storage medium and computer device for processing tender evaluation data. Background Art

[0002] In the process of tender evaluation, various types of data of the bidders need to be managed and processed to provide a scientific decision-making basis for the tender evaluation process.

[0003] Currently, in the tender process, the original data of the bidders is usually evaluated manually to obtain the tender evaluation basis. However, this manual processing method will lead to low efficiency of data processing. At the same time, affected by subjective factors, the data processing will not be objective enough, that is, the accuracy of data processing is low, resulting in the final evaluation result not being objective and accurate enough. Summary of the Invention

[0004] The present invention provides a method, device, storage medium and computer device for processing tender evaluation data, mainly aiming at improving the processing efficiency and accuracy of tender evaluation data, and making the evaluation result more objective and accurate.

[0005] According to the first aspect of the present invention, a method for processing tender evaluation data is provided, including:

[0006] Responding to a tender signal of a target tender project, obtaining the data to be evaluated of the bidder, and determining a plurality of evaluation indicators of the target tender project;

[0007] Classifying the data to be evaluated based on each of the evaluation indicators to obtain the data to be evaluated under different categories;

[0008] Inputting the data to be evaluated under each category into a preset scoring model for scoring to obtain a predicted score under each evaluation indicator;

[0009] Determining an evaluation terminal corresponding to each evaluation indicator, and respectively sending the data to be evaluated corresponding to each evaluation indicator to the corresponding evaluation terminal, so as to determine a terminal score under each evaluation indicator through each evaluation terminal based on the corresponding data to be evaluated;

[0010] Receiving the terminal score of each evaluation indicator, and determining the evaluation result of the bidder based on the predicted score and the terminal score under each evaluation indicator.

[0011] Optionally, the classifying the data to be evaluated based on each of the evaluation indicators to obtain the data to be evaluated under different categories includes:

[0012] Determine the centroid vector of the cluster corresponding to each of the review indicators, and determine the review feature vector corresponding to each of the to-be-reviewed data;

[0013] Based on the review feature vector and the centroid vector, calculate the distance between each of the to-be-reviewed data and the centroid of each cluster, and divide each of the to-be-reviewed data into each of the clusters based on the distance;

[0014] Based on the review feature vector corresponding to the to-be-reviewed data in each of the clusters, determining an updated centroid vector corresponding to each of the clusters;

[0015] Based on the updated centroid vector, each of the to-be-reviewed data is re-divided into each of the clusters until the updated centroid vector does not change, and the to-be-reviewed data finally divided into each of the clusters is determined as to-be-reviewed data under different categories.

[0016] Optionally, before classifying the data to be reviewed based on each of the review indicators to obtain the data to be reviewed in different categories, the method further includes:

[0017] Based on the identifiers of the data to be reviewed, determine the same data in the data to be reviewed, and based on the timestamp information of each data in the same data, determine the redundant data in the same data, and remove the redundant data to obtain the processed data to be reviewed;

[0018] Determine the data density in a preset neighborhood corresponding to each data in the processed data to be reviewed, and determine the core data and non-core data in each data based on the data density;

[0019] Taking any core data in the core data as target core data, clustering the remaining core data with the target core data as the cluster center to obtain core data under different clustering categories;

[0020] Taking any non-core data in the non-core data as target non-core data, determining the reference core data closest to the non-core data in the core data under each clustering category, and judging whether the distance between the non-core data and the reference core data is less than a preset distance threshold;

[0021] If the distance between the non-core data and the reference core data is less than the preset distance threshold, the non-core data is classified into the cluster category to which the reference core data belongs; otherwise, the non-core data is determined as abnormal data;

[0022] Eliminating the abnormal data from the processed data to be reviewed to obtain the cleaned data to be reviewed;

[0023] Classify the data to be reviewed based on each of the review metrics to obtain the data to be reviewed under different categories, including:

[0024] Classify the cleaned data to be reviewed of the bidders based on each of the review metrics to obtain the data to be reviewed under different categories.

[0025] Optionally, each review terminal determines the terminal score under each review metric based on the corresponding data to be reviewed, including:

[0026] In the case of multiple bidders, any review metric in each review metric is respectively used as a target review metric, and multiple comparison items under the target review metric are determined;

[0027] Determine the data to be compared under each comparison item in the data to be reviewed corresponding to the target review metric;

[0028] Standardize the data to be compared of each bidder under each comparison item to obtain the standardized data to be compared;

[0029] Generate a difference comparison graph of each bidder under the same comparison item based on the standardized data to be compared of each bidder under each comparison item, wherein the horizontal axis of the difference comparison graph represents each bidder, and the vertical axis represents the standardized data to be compared of each bidder under the same comparison item;

[0030] Send each difference comparison graph corresponding to each review metric to the corresponding review terminal, so that each review terminal determines the terminal score under each review metric based on the corresponding difference comparison graph.

[0031] Optionally, before obtaining the data to be reviewed of the bidders, the method further includes:

[0032] Determine the affiliated enterprises of the enterprise to which the target tender project belongs, and determine multiple network servers corresponding to each affiliated enterprise;

[0033] Obtain various data of multiple cooperative bidders in each network server, encrypt the various data, and store the encrypted various data in the cache;

[0034] The obtaining of the data to be reviewed of the bidders includes:

[0035] Based on the project category carried in the tender signal, obtain the encrypted data of the tenderer under the project category in the cache, and obtain the data submitted by the tenderer during the current tendering process, and form the data to be reviewed with the encrypted data and the submitted data.

[0036] Optionally, determining the review result of the tenderer based on the predicted score and the terminal score under each review index includes:

[0037] Determine the index weight corresponding to each review index respectively, and determine the score weight corresponding to the predicted score and the terminal score respectively;

[0038] Based on the score weight, add the predicted score and the terminal score under each review index to obtain the index score under each review index;

[0039] Based on the index weight, add the index scores under each review index to obtain the comprehensive score of the tenderer, and determine the review result of the tenderer based on the comprehensive score.

[0040] Optionally, before inputting the data to be reviewed under each category into a preset scoring model for scoring to obtain the predicted score under each review index, the method further includes:

[0041] Construct multiple initial scoring models, and obtain a sample data set, where the sample data set includes sample review data under different indexes with label information, and the label information is the actual score corresponding to different indexes;

[0042] Based on the number of models of the initial scoring models, divide the sample data set into multiple groups of training data and multiple groups of test data;

[0043] Use each group of training data to train the corresponding initial scoring model, and use each group of test data to test the corresponding trained initial scoring model, and determine the initial scoring model that meets the test conditions as the preset scoring model.

[0044] According to the second aspect of the present invention, there is provided a processing device for tender review data, including:

[0045] An acquisition unit, configured to obtain the data to be reviewed of the tenderer in response to the tender signal of the target tender project, and determine multiple review indexes of the target tender project;

[0046] A classification unit, configured to classify the data to be reviewed based on each review index to obtain the data to be reviewed under different categories;

[0047] A first scoring unit, configured to respectively input the data to be reviewed under each of the categories into a preset scoring model for scoring, so as to obtain a predicted score under each of the review indicators;

[0048] A second scoring unit, configured to determine a review terminal corresponding to each of the review indicators, and respectively send the data to be reviewed corresponding to each of the review indicators to the corresponding review terminal, so as to determine, by each of the review terminals based on the corresponding data to be reviewed, a terminal score under each of the review indicators;

[0049] A determination unit, configured to receive the terminal scores of each of the review indicators, and determine a review result of the tenderer based on the predicted scores and the terminal scores under each of the review indicators.

[0050] According to a third aspect of the present invention, there is provided a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned method for processing tender review data is implemented.

[0051] According to a fourth aspect of the present invention, there is provided a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, the above-mentioned method for processing tender review data is implemented.

[0052] According to a method, device, storage medium, and computer device for processing tender review data provided by the present invention, compared with the current method of manually processing the data of tenderers to obtain the basis for tender review, the present invention classifies the data to be reviewed of tenderers through review indicators, and first respectively inputs the data to be reviewed under each category into a preset scoring model for scoring different review indicators. At the same time, the data to be reviewed of different categories are respectively sent to the corresponding review terminals for scoring different review indicators. Finally, according to the scores predicted by the model and the scores given by the review terminals, the review result of the tenderer is determined. Thus, by pre-classifying the data to be reviewed according to the review indicators, when predicting the scores under different review indicators, the time and computing resources wasted by analyzing the data to be reviewed that are not related to the review indicators can be avoided, thereby improving the efficiency of score determination, saving computing resources, and further improving the efficiency of review result determination. At the same time, the present invention comprehensively analyzes the scores respectively determined by the review terminals and the model to determine the final review result, which can avoid the limitations of a single evaluation method and the problems of low processing efficiency and low processing accuracy caused by solely relying on manual processing of the review process. Therefore, the present invention can improve the accuracy and objectivity of the tender review process, and further improve the accuracy of the review result. Description of the Drawings

[0053] The accompanying drawings described herein are used to provide a further understanding of the present invention and form a part of this application. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0054] Figure 1 A flowchart of a method for processing tender evaluation data provided by an embodiment of the present invention is shown;

[0055] Figure 2 A schematic diagram of a network architecture for integrated information sharing provided by an embodiment of the present invention is shown;

[0056] Figure 3 A flowchart of another method for processing tender evaluation data provided by an embodiment of the present invention is shown;

[0057] Figure 4 A schematic diagram of the structure of a device for processing tender evaluation data provided by an embodiment of the present invention is shown;

[0058] Figure 5 A schematic diagram of the physical structure of a computer device provided by an embodiment of the present invention is shown. Detailed implementation manners

[0059] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments. It should be noted that, without conflict, the embodiments in this application and the features in the embodiments may be combined with each other.

[0060] Currently, the method of manually evaluating and processing the original data of bidders to obtain the basis for tender evaluation leads to low efficiency in data processing. At the same time, affected by subjective human factors, the final evaluation results are not objective and accurate enough.

[0061] To solve the above problems, an embodiment of the present invention provides a method for processing tender evaluation data, as Figure 1 shown, the method includes:

[0062] 101. In response to the tender signal of the target tender project, obtain the data to be evaluated of the bidder and determine multiple evaluation indicators of the target tender project.

[0063] Among them, the target tender project can be any project, such as the tender procurement project of the distribution network agreement inventory materials of power grid enterprises, the tender project of power grid infrastructure projects, etc.; the data to be reviewed refers to the qualification data of the bidders, company performance, company reputation, company profile data, performance and cases, various certificates and reports, technical solutions, technical strength (such as the R & D team of the bidders, the number of technical personnel, the number of technical patents, the technical innovation ability, etc.), technical quality, technical training, after-sales support, tender offer, cost composition, profit margin, market competitiveness, price concessions and discounts, management system, team strength, cooperation experience, etc.; the multiple review indicators include: technical indicators, commercial indicators, price indicators, service and support indicators, delivery period indicators, team strength indicators, etc.; the review indicators are preset according to the actual review requirements.

[0064] For the embodiments of the present invention, an intranet tender review data management system is proposed. The intranet tender review data management system can realize data collection and storage: the system can automatically collect various data to be reviewed required for review, such as prequalification results, performance evaluations, technical solutions, and bad behaviors, from each digital information system of the enterprise and each tender review link, and store them in a temporary storage environment. Data processing and distribution: the system can process the collected data according to the needs of tender review work, match the corresponding bidders according to different review indicators and review processes. Data analysis and decision support: the system provides a variety of data analysis and decision support functions to help better understand and utilize the data, and improve the accuracy and efficiency of the review.

[0065] Specifically, such as Figure 2The figure shows a network architecture for integrated information sharing. This network architecture can rely on the bid evaluation intranet environment of each branch of an enterprise and be applied in the tender evaluation stage. Based on a distributed storage architecture, it integrates existing servers, internal platform systems, and bid evaluation terminal devices to form a network architecture for integrated information sharing. After a tender project is determined, the information data of multiple systems incorporated into the information integration network architecture can be integrated and processed according to requirements, including deduplication, verification, comparison of association relationships, encryption, transmission, storage, and classification, distribution, preprocessing, etc. according to business requirements. For example, for the tender procurement of distribution network protocol inventory materials, the to-be-evaluated data in the prequalification library of an enterprise's e-commerce platform can be classified and screened according to dimensions such as bidders, material categories, and evaluation indicators and synchronized to the bid evaluation intranet environment, and data extraction and processing can be performed according to the needs of bid evaluation to obtain the evaluation results. During the evaluation, various tools can be used to access the distributed storage server through various system interfaces to obtain relevant evaluation data, and the data can be presented in a structured manner on the tools, intuitively and conveniently presenting various to-be-evaluated data of bidders, providing objective and true data as a basis for the evaluation work. In the scoring session, the process data and the final scores are centrally stored and can be shared through interfaces with other required systems to achieve data unity. After the evaluation work is completed, the overall data can be backed up, stored, or deleted according to user needs, realizing full business process traceability and ensuring data security to the greatest extent. The embodiments of the present invention use scalable data storage and processing technologies to improve the performance and operation processing speed of existing storage devices. At the same time, by applying a distributed computing framework, large-scale data sets are processed on multiple computers, maximizing the efficiency of the original equipment and supporting different types of data processing tasks such as batch processing, stream processing, and graph processing. A new type of database is used to achieve distributed storage, and data processing, reading, and transmission efficiency are optimized through data partitioning, replication, and compression technologies. The information barrier between the internal and external networks is broken through the optimized and upgraded interfaces, while taking into account the security requirements of the internal network. The required data is downloaded and saved in an independent internal network space at one time, reducing problems such as poor timeliness, vulnerability to attacks, and a large amount of redundant data caused by multiple data interactions. Information sharing is formed among multiple systems in the bid evaluation intranet. Multiple systems can be processed at one time and can be commonly used and updated in real time, facilitating managers to monitor the operation conditions of multiple systems.

[0066] 102. Classify the to-be-evaluated data based on each evaluation indicator to obtain the to-be-evaluated data under different categories.

[0067] For the embodiments of the present invention, after obtaining the data to be evaluated of the bidders and determining the evaluation indicators, the data to be evaluated can be classified according to the evaluation indicators, that is, each data to be evaluated is classified into the categories of different evaluation indicators. For example, the data to be evaluated related to technical standards such as technical solutions, technical quality, and technical advancement are classified into the category of technical indicators, and the data to be evaluated related to business standards such as company qualifications, company performance, and company reputation are classified into the category of business indicators. Then, according to the data to be evaluated under different evaluation indicator categories, the evaluation results corresponding to different evaluation indicators are determined. By classifying the data to be evaluated, the embodiments of the present invention can conduct targeted evaluations for each evaluation indicator, avoid the dilemma of blindly searching for relevant information in a large amount of data, save evaluation resources, and significantly improve the evaluation efficiency.

[0068] 103. Input the data to be evaluated under each category into a preset scoring model for scoring to obtain the predicted scores under each evaluation indicator.

[0069] For the embodiments of the present invention, the preset scoring model is trained based on a sample data set. The sample data set includes the data to be evaluated under different evaluation indicators with annotation information, and the annotation information is the actual scores corresponding to different evaluation indicators. Each category is the category corresponding to each evaluation indicator. For example, input the data to be evaluated under the technical indicators into the preset scoring model, and the preset scoring model can output the technical scores corresponding to the technical indicators. Input the data to be evaluated under the price indicators into the preset scoring model, and the preset scoring model can output the price scores corresponding to the price indicators.

[0070] 104. Determine the evaluation terminals corresponding to each evaluation indicator, and send the data to be evaluated corresponding to each evaluation indicator to the corresponding evaluation terminals respectively, so as to determine the terminal scores under each evaluation indicator through each evaluation terminal based on the corresponding data to be evaluated.

[0071] For the embodiments of the present invention, each evaluation terminal corresponds to an evaluation expert. For example, there is a technical expert in the evaluation terminal corresponding to the technical indicators, a business expert in the evaluation terminal corresponding to the business indicators, and a price expert in the evaluation terminal corresponding to the price indicators. Send the data to be evaluated under the technical indicators to the evaluation terminal where the technical expert is located, and the technical expert analyzes each data to be evaluated to obtain the technical scores under the technical indicators. Send the data to be evaluated under the business indicators to the evaluation terminal where the business expert is located, and the business expert analyzes each data to be evaluated to obtain the business scores under the business indicators. Thus, the terminal scores determined by the evaluation terminal for each evaluation indicator can be obtained.

[0072] 105. Receive the terminal scores for each evaluation indicator, and determine the evaluation result of the tenderer based on the predicted scores and the terminal scores under each evaluation indicator.

[0073] For the embodiments of the present invention, after each evaluation terminal determines the terminal scores under the corresponding evaluation indicators, it will send back each terminal score. After receiving each terminal score, a comprehensive analysis is performed on the predicted scores and terminal scores under each evaluation indicator to obtain the total tender score of the tenderer. Finally, based on the total tender score, the tender result of the tenderer is determined. Thus, the embodiments of the present invention determine the final evaluation result by comprehensively analyzing the scores determined by the evaluation terminal and the model respectively, which can avoid the limitations of a single evaluation method and also avoid the problems of low processing efficiency and low processing accuracy caused by solely relying on manual processing of the evaluation process. Therefore, the present invention can improve the accuracy and objectivity of the tender evaluation process, and further improve the accuracy of the evaluation result.

[0074] According to a method for processing tender evaluation data provided by the present invention, compared with the current method of manually processing the data of tenderers to obtain the basis for tender evaluation, the present invention classifies the data to be evaluated of tenderers through evaluation indicators, and first inputs the data to be evaluated under each category into a preset scoring model for scoring different evaluation indicators. At the same time, the data to be evaluated in different categories are separately sent to the corresponding evaluation terminals for scoring different evaluation indicators. Finally, based on the scores predicted by the model and the scores given by the evaluation terminals, the evaluation result of the tenderer is determined. Thus, by pre-classifying the data to be evaluated according to evaluation indicators, when predicting the scores under different evaluation indicators, it is possible to avoid wasting time and computing resources in analyzing the data to be evaluated that is not relevant to the evaluation indicators, thereby improving the efficiency of determining scores, saving computing resources, and further improving the efficiency of determining the evaluation result. At the same time, the present invention determines the final evaluation result by comprehensively analyzing the scores determined by the evaluation terminal and the model respectively, which can avoid the limitations of a single evaluation method and also avoid the problems of low processing efficiency and low processing accuracy caused by solely relying on manual processing of the evaluation process. Therefore, the present invention can improve the accuracy and objectivity of the tender evaluation process, and further improve the accuracy of the evaluation result.

[0075] Further, in order to better illustrate the above process of processing tender evaluation data, as a refinement and extension of the above embodiments, the embodiments of the present invention provide another method for processing tender evaluation data, as Figure 3 shown, the method includes:

[0076] 201. In response to the tender signal of the target tender project, obtain the data to be evaluated of the tenderer and determine multiple evaluation indicators of the target tender project.

[0077] For the embodiments of the present invention, in order to increase the richness of the data to be reviewed and thus improve the review accuracy, all previous bidding data can be stored so that when it is necessary to review a certain bidder, the data of the bidder in the previous bidding process can be used for reference. Based on this, the method includes: determining the affiliated enterprises of the enterprise to which the target bidding project belongs, and determining a plurality of network servers corresponding to each of the affiliated enterprises; obtaining various types of data of a plurality of cooperative bidders in each of the network servers, encrypting the various types of data, and storing the encrypted various types of data in a cache.

[0078] Specifically, each branch company under the bidding enterprise has one or more bid evaluation bases, and each bid evaluation base also has a plurality of network servers. Through the construction of a distributed storage architecture, independent servers are combined into a temporary storage environment, which is uniformly managed and operated in the internal bid evaluation network. With the help of a dedicated interface between the bid evaluation base and the external network, on the premise of ensuring information security, data information related to bid evaluation is downloaded from multiple external network systems to the temporary storage environment (cache) at one time, and these data are sorted and saved according to project categories and bidders; when review is needed, relevant data can be quickly read from the cache through various bid evaluation tools, and for some data that needs to be calculated and analyzed, the overall servers of the temporary storage environment can also be used for quick calculation. Further, using the distributed storage architecture, the servers of each bid evaluation base are carefully combined and connected in series to build a dedicated temporary storage environment. Then, comprehensive optimization is carried out for the processor, memory, network, etc., and technical innovation operations such as upgrading the operating system and database structure are performed to achieve efficient integration and optimization of the original independent devices. When the bid evaluation work is officially started, make full use of the original data interface to obtain relevant various types of data from numerous internal and external network platforms according to actual needs at one time, and accurately store them in the already built temporary storage environment. When data needs to be read subsequently, directly access the database of the temporary storage environment uniformly, which greatly reduces the data interaction between multiple systems, thus significantly improving the system performance and reducing the load. At the same time, transfer the work of processing data originally responsible for by a single system to the entire temporary storage environment for collaborative processing. For the data generated during the bid evaluation process, first perform strict encryption processing, and then uniformly store it in the cache. By encrypting and storing the data, the embodiments of the present invention can ensure the security of the data.

[0079] Further, according to the project category carried in the tender invitation signal, encrypted data of the tenderer under this project category is obtained from the cache, and data submitted by the tenderer during the current tendering process is obtained. The data to be reviewed is composed of the encrypted data and the submitted data. In the embodiment of the present invention, by obtaining the data in the previous tendering process of the enterprise and comprehensively analyzing the previous data and the data submitted during the current tendering process, richer data support is provided for the review process, and the review accuracy is increased.

[0080] 202. Based on the identifiers of each data to be reviewed, the same-kind data is determined in the data to be reviewed. Based on the timestamp information of each data in the same-kind data, redundant data is determined in the same-kind data, and the redundant data is removed to obtain the processed data to be reviewed.

[0081] Among them, the identifier can be the ID, name, code, etc. of the data to be reviewed; the timestamp information is the information recording the generation or update time of the data to be reviewed.

[0082] Specifically, based on the identifier of the data, the data to be reviewed is classified into the same-kind data. The same-kind data refers to the data with the same identifier, and each data in the same-kind data is the same data at different time points. Based on the timestamp, the same-kind data is sorted, and it is determined which data is old. The old data (redundant data) generally refers to the data item with an earlier timestamp in the same-kind data. It is also possible to set a time threshold according to specific business requirements, and regard the data with a timestamp earlier than the threshold as old data. That is, through the comparison of the data to be reviewed, the latest data is retained according to the data timeliness (timestamp information) sorting, and the old data is deleted. Finally, the collected information is screened, verified, and changed into structured data storage according to the fields in the structured data table of the review rules for different projects.

[0083] 203. Abnormal data is determined in the processed data to be reviewed, and the abnormal data in the processed data to be reviewed is removed to obtain the cleaned data to be reviewed.

[0084] For the embodiments of the present invention, in order to avoid the interference of abnormal data in the data to be reviewed on the review result and improve the data quality of the data to be reviewed, it is first necessary to determine the abnormal data in the data to be reviewed. Based on this, step 203 specifically includes: determining the data density within the preset neighborhood corresponding to each data in the processed data to be reviewed, and based on the data density, determining the core data and non-core data among each data; taking any core data in the core data as a target core data respectively, and using the target core data as a clustering center to cluster the remaining core data to obtain the core data under different clustering categories; taking any non-core data in the non-core data as a target non-core data respectively, determining the reference core data closest to the non-core data among the core data under each clustering category, and judging whether the distance between the non-core data and the reference core data is less than a preset distance threshold; if the distance between the non-core data and the reference core data is less than the preset distance threshold, then classifying the non-core data into the clustering category to which the reference core data belongs, otherwise, determining the non-core data as abnormal data.

[0085] Among them, the preset neighborhood is set according to actual needs, and the preset distance threshold is set according to actual needs. The data density can specifically be the data volume. Specifically, if the data density within the preset neighborhood corresponding to a certain data in the data to be reviewed is greater than the preset density threshold, then this data is determined as core data. On the contrary, if the data density within the preset neighborhood corresponding to a certain data is less than or equal to the preset density threshold, then this data is determined as non-core data. Then, each core data is used as a clustering center respectively to cluster the remaining core data to obtain different clustering categories. Then, it is determined whether each non-core data can be clustered into the above-mentioned clustering categories. Finally, the non-core data that cannot be clustered into all the above-mentioned clustering categories is determined as abnormal data. Further, the abnormal data in the data to be reviewed is deleted to obtain the cleaned data to be reviewed. The embodiments of the present invention can improve the data quality, avoid the computing resources wasted by useless data parameter operations, and also avoid abnormal data participating in the operation from disturbing the review result by cleaning the data to be reviewed. That is, the embodiments of the present invention can improve the data quality of the data to be reviewed, improve the review effect, and reduce the computing cost by performing outlier detection and elimination on the data to be reviewed.

[0086] In another embodiment of the present invention, anomaly detection for the data to be reviewed can also be determined in the following manner: performing a form conversion on the data to be reviewed to obtain the data to be reviewed of numerical type; determining the median corresponding to the data to be reviewed of numerical type, and determining a first intermediate value between the minimum value in the data to be reviewed of numerical type and the median, and determining the first intermediate value as the first quartile corresponding to the data to be reviewed of numerical type; determining a second intermediate value between the maximum value in the data to be reviewed of numerical type and the median, and determining the second intermediate value as the third quartile corresponding to the data to be reviewed of numerical type; calculating the distance between the first quartile and the third quartile, and determining the distance as the interquartile range corresponding to the data to be reviewed of numerical type; calculating the lower limit value of anomaly detection corresponding to the data to be reviewed of numerical type according to the first quartile and the interquartile range; calculating the upper limit value of anomaly detection corresponding to the data to be reviewed of numerical type according to the third quartile and the interquartile range; in the data to be reviewed of numerical type, determining the data outside the lower limit value of anomaly detection to the upper limit value of anomaly detection as abnormal data.

[0087] Specifically, the data to be reviewed are uniquely encoded, that is, a new virtual variable is created for each data to be reviewed, so as to obtain the data to be reviewed of the numerical type. Then, the median Q2 and the middle value of the data to be reviewed of the numerical type are first determined, and according to the median, the first quartile Q1 corresponding to the data to be reviewed of the numerical type is determined, that is, the middle number between the minimum value and the median in the data to be reviewed of the numerical type; then, according to the median, the third quartile Q3 corresponding to the data to be reviewed of the numerical type is determined, that is, the middle number between the median and the maximum value of the data to be reviewed of the numerical type; and the interquartile range IQR corresponding to the data to be reviewed of the numerical type is determined, that is, the distance between the first quartile and the third quartile; then, the upper limit value upper=Q3+1.5*IQR and the lower limit value lower=Q1-1.5*IQR corresponding to the data to be reviewed of the numerical type are calculated, and finally, the data other than lower-upper is determined in the data to be reviewed of the numerical type, and the data other than lower-upper is determined as abnormal data. Then, the abnormal data in the processed data to be reviewed is eliminated to obtain the cleaned data to be reviewed. It should be noted that the method for determining abnormal data is not limited to the above method, and other abnormality detection methods may also be used.

[0088] 204. Based on each review indicator, the cleaned data to be reviewed is classified to obtain data to be reviewed in different categories.

[0089] For the embodiments of the present invention, after cleaning the data to be reviewed, it is also necessary to classify the cleaned data to be reviewed. Based on this, step 204 specifically includes: determining the centroid vectors of the clusters corresponding to each review index, and determining the review feature vectors corresponding to each data to be reviewed; based on the review feature vectors and the centroid vectors, calculating the distances between each data to be reviewed and the centroid of each cluster, and based on the distances, dividing each data to be reviewed into each cluster; based on the review feature vectors corresponding to the data to be reviewed in each cluster, determining the updated centroid vectors corresponding to each cluster; based on the updated centroid vectors, re-dividing each data to be reviewed into each cluster until the updated centroid vectors do not change, and determining the data to be reviewed finally divided into each cluster as the data to be reviewed under different categories.

[0090] Specifically, first, use methods such as word embedding to determine the review feature vectors corresponding to each data to be reviewed, and determine the centroid vectors of the clusters corresponding to each review index. For the review feature vectors corresponding to each data to be reviewed respectively, calculate the distances from each review feature vector to each centroid vector, and according to the distances, allocate each data to be reviewed to the cluster corresponding to the centroid vector with the closest distance. Then, for each cluster, recalculate the centroid of each cluster and its corresponding centroid vector, and re-divide each data to be reviewed into different clusters, so as to continuously divide each data to be reviewed until the position of the centroid does not change, that is, the centroid vector does not change. Finally, determine the data to be reviewed divided into different clusters as the data to be reviewed under different categories.

[0091] Further, in order to improve the classification accuracy, it is first necessary to accurately determine the centroid of the cluster corresponding to each review index. Based on this, the method includes: determining the number of indicators corresponding to the review index; determining the data density within the preset neighborhood corresponding to each data to be reviewed, and determining the density mean corresponding to each data density; based on the density mean, dividing each data to be reviewed into a high-density set and a low-density set; determining the high-density data to be reviewed with a data density greater than the preset density threshold in the data to be reviewed in the high-density set, and determining a preset number of central data to be reviewed with a mutual distance greater than the preset distance threshold among the high-density data to be reviewed, and using the central data to be reviewed as the centroid corresponding to each review index cluster respectively. The preset number is the same as the number of indicators.

[0092] Among them, the preset neighborhood is set according to actual requirements, the preset density threshold is set according to actual requirements, and the preset distance is set according to actual requirements. Specifically, the data density within the preset neighborhood corresponding to each data to be reviewed is determined respectively. For example, the data density corresponding to the data to be reviewed 1 is 3, the data density corresponding to the data to be reviewed 2 is 4, the data density corresponding to the data to be reviewed 3 is 5, the data density corresponding to the data to be reviewed 4 is 9, the data density corresponding to the data to be reviewed 5 is 9, and the data density corresponding to the data to be reviewed 6 is 10. Then the density average value is 6.7. And the data to be reviewed 4, the data to be reviewed 5, and the data to be reviewed 6 with data density greater than 6.7 are classified into the high-density set, and the data to be reviewed 1, the data to be reviewed 2, and the data to be reviewed 3 are classified into the low-density set. If the preset density threshold is 5 and the preset quantity is 2, then the data with a mutual distance greater than the preset threshold are finally determined to be the data to be reviewed 4 and the data to be reviewed 6. Finally, the data to be reviewed 4 and the data to be reviewed 6 are respectively determined as the centroids corresponding to each cluster. The method according to the embodiment of the present invention determines the centroid of the cluster by means of data density and mutual dissimilarity. By considering the data density, the natural distribution characteristics of the data can be more accurately reflected. The high-density area usually corresponds to the core part of the data set, while the low-density area may contain noise or outliers. Therefore, selecting the initial centroid according to the data density helps to place the centroid at the core position of the data, thereby improving the accuracy of centroid determination and further improving the classification accuracy.

[0093] 205. Input the data to be reviewed under each category into the preset scoring model respectively for scoring, and obtain the predicted scores under each evaluation index.

[0094] For the embodiment of the present invention, in order to improve the prediction accuracy of the preset scoring model, it is first necessary to train and construct the preset scoring model. Based on this, the method includes: constructing a plurality of initial scoring models, and obtaining a sample data set, wherein the sample data set includes sample review data under different indexes with label information, and the label information is the actual scores corresponding to different indexes; dividing the sample data set into multiple groups of training data and multiple groups of test data based on the number of the initial scoring models; training each initial scoring model with the corresponding training data, and testing each trained initial scoring model with the corresponding test data, and determining the initial scoring model that meets the test conditions as the preset scoring model.

[0095] Specifically, during the process of training the initial scoring model using training data, the sample review data under each indicator is used as input data, and the actual score corresponding to each indicator is used as output data. During the process of testing the trained initial scoring model, the sample review data under each indicator in the test data is used as input data, and the predicted score under each indicator is output through the trained initial scoring model. According to the actual score and the predicted score corresponding to each indicator, the prediction accuracy of the trained initial scoring model is determined. Finally, the initial scoring model with the highest prediction accuracy is determined as the preset scoring model.

[0096] 206. Determine the review terminal corresponding to each review indicator, and separately send the data to be reviewed corresponding to each review indicator to the corresponding review terminal, so as to determine the terminal score under each review indicator through each review terminal based on the corresponding data to be reviewed.

[0097] For the embodiments of the present invention, in order to avoid the limitations of review and thus improve the review accuracy, in addition to using the preset scoring model to score the bidders under different review indicators, it is also necessary to use the experts of the review terminal to score the bidders under different review indicators. Based on this, step 206 specifically includes: in the case of multiple bidders, any review indicator in each of the review indicators is respectively used as a target review indicator, and multiple comparison items under the target review indicator are determined; the data to be compared under each comparison item is determined in the data to be reviewed corresponding to the target review indicator; the data to be compared under each comparison item for each bidder is standardized to obtain the standardized data to be compared; based on the standardized data to be compared under each comparison item for each bidder, a difference comparison graph of each bidder under the same comparison item is generated, where the horizontal axis of the difference comparison graph represents each bidder, and the vertical axis represents the standardized data to be compared of each bidder under the same comparison item; each difference comparison graph corresponding to each review indicator is sent to the corresponding review terminal, so as to determine the terminal score under each review indicator through each review terminal based on the corresponding difference comparison graph.

[0098] Specifically, the comparison items for different indicators are as follows. For technical indicators, the corresponding comparison items include supply performance, bid response, component materials, etc.; for business indicators, the corresponding comparison items include company qualifications, company performance, contract execution ability, etc.

[0099] Specifically, the standardization process refers to processing the data to be reviewed of each bidder into data with a unified format, unit, etc. Classify the data to be reviewed under each evaluation index according to the comparison items to obtain the data to be reviewed under different comparison items. Generate a difference comparison graph based on the data to be reviewed of different bidders under the same comparison item. For example, for the business index, a difference comparison graph corresponding to the company's qualifications, a difference comparison graph corresponding to the company's performance, and a difference comparison graph corresponding to the contract execution ability will be generated. In the above manner, difference comparison graphs corresponding to different comparison items under different evaluation indexes can be generated. The difference comparison graph can be any one of a bar graph, a bar chart, a line graph, etc. The embodiment of the present invention does not specifically limit the form of the difference comparison graph. Then, send each difference comparison graph corresponding to different evaluation indexes to the corresponding evaluation terminal. For example, send each difference comparison graph under the technical index to the evaluation terminal corresponding to the technical expert, and send each difference comparison graph under the business index to the evaluation terminal corresponding to the business expert. Further, after each evaluation terminal receives the corresponding difference comparison graph, it will combine the difference comparison graph to score the corresponding evaluation index. Since the difference comparison graph graphically shows the data differences of the bidders in each evaluation dimension (comparison item), the evaluation terminal can quickly capture key information, avoid getting stuck in cumbersome data reading, and at the same time can easily make a horizontal comparison of the performances of different bidders, so as to more accurately evaluate the advantages and disadvantages of each bidder, and further improve the evaluation efficiency and evaluation accuracy.

[0100] 207. Receive the terminal scores of each evaluation index, and based on the predicted scores and terminal scores under each evaluation index, determine the evaluation result of the bidder.

[0101] For the embodiment of the present invention, after predicting the predicted scores under each evaluation index using the preset scoring model and the evaluation terminal gives the terminal scores for each evaluation index, it is necessary to comprehensively analyze the above predicted scores and terminal scores to determine the evaluation result. Based on this, step 207 specifically includes: respectively determining the index weights corresponding to each of the evaluation indexes, and determining the scoring weights corresponding to the predicted scores and the terminal scores respectively; based on the scoring weights, add the predicted scores and the terminal scores under each of the evaluation indexes to obtain the index scores under each of the evaluation indexes; based on the index weights, add the index scores under each of the evaluation indexes to obtain the comprehensive score of the bidder, and based on the comprehensive score, determine the evaluation result of the bidder.

[0102] Among them, the index weights and scoring weights are set according to actual needs. Specifically, if the index weight corresponding to evaluation index 1 is λ 1 , the index weight corresponding to evaluation index 2 is λ 2 , the index weight corresponding to evaluation index 3 is λ3 The scoring weight corresponding to the predicted score is ω 1 The scoring weight corresponding to the terminal score is ω 2 For review index 1, the predicted score is a and the terminal score is b; for review index 2, the predicted score is c and the terminal score is d; for review index 3, the predicted score is e and the terminal score is f. Then the comprehensive score y = λ 1 (ω 1 a + ω 2 b) + λ 2 (ω 1 c + ω 2 d) + λ 3 (ω 1 e + ω 2 f). Finally, among all the bidders, the target bidder corresponding to the highest comprehensive score is determined as the winning bidder. In the embodiments of the present invention, by setting different weights for different review indexes and review methods, the review process can be ensured to be more targeted at the specific requirements and characteristics of the project. The weight allocation can reflect the importance of each review index and review method in the overall evaluation, making the review results more accurate.

[0103] According to another method for processing tender review data provided by the present invention, compared with the current method of manually processing the data of bidders to obtain the tender review basis, the present invention classifies the data to be reviewed of bidders through review indexes, and first inputs the data to be reviewed under each category into a preset scoring model for scoring different review indexes. At the same time, the data to be reviewed of different categories are separately sent to the corresponding review terminals for scoring different review indexes. Finally, according to the scores predicted by the model and the scores given by the review terminals, the review results of the bidders are determined. Thus, by pre-classifying the data to be reviewed according to the review indexes, when predicting the scores under different review indexes, the time and computing resources wasted by analyzing the data to be reviewed that are not related to the review indexes can be avoided, thereby improving the efficiency of score determination, saving computing resources, and further improving the efficiency of determining the review results. At the same time, the present invention determines the final review results by comprehensively analyzing the scores determined by the review terminals and the model respectively, which can avoid the limitations of a single evaluation method and the problems of low processing efficiency and low processing accuracy caused by solely relying on manual processing of the review process. Therefore, the present invention can improve the accuracy and objectivity of the tender review process, and further improve the accuracy of the review results.

[0104] Further, as a Figure 1 specific implementation, the embodiments of the present invention provide a device for processing tender review data, as Figure 4 shown. The device includes: an acquisition unit 31, a classification unit 32, a first scoring unit 33, a second scoring unit 34, and a determination unit 35.

[0105] The obtaining unit 31 can be configured to obtain the data to be reviewed of the bidders in response to the bidding signal of the target bidding project, and determine multiple evaluation indicators of the target bidding project.

[0106] The classification unit 32 can be configured to classify the data to be reviewed based on each of the evaluation indicators, and obtain the data to be reviewed under different categories.

[0107] The first scoring unit 33 can be configured to input the data to be reviewed under each category into a preset scoring model for scoring respectively, and obtain the predicted scores under each of the evaluation indicators.

[0108] The second scoring unit 34 can be configured to determine the evaluation terminals corresponding to each of the evaluation indicators, and send the data to be reviewed corresponding to each of the evaluation indicators to the corresponding evaluation terminals respectively, so as to determine the terminal scores under each of the evaluation indicators through each of the evaluation terminals based on the corresponding data to be reviewed.

[0109] The determining unit 35 can be configured to receive the terminal scores of each of the evaluation indicators, and determine the evaluation result of the bidder based on the predicted scores and the terminal scores under each of the evaluation indicators.

[0110] In a specific application scenario, in order to classify the data to be reviewed, the classification unit 32 includes a first determining module 321 and a dividing module 322.

[0111] The first determining module 321 can be configured to determine the centroid vectors of the clusters corresponding to each of the evaluation indicators, and determine the evaluation feature vectors corresponding to each of the data to be reviewed.

[0112] The dividing module 322 can be configured to calculate the distances between each of the data to be reviewed and the centroid of each cluster based on the evaluation feature vectors and the centroid vectors, and divide each of the data to be reviewed into each cluster based on the distances.

[0113] The first determining module 321 can also be configured to determine the updated centroid vectors corresponding to each of the clusters based on the evaluation feature vectors corresponding to the data to be reviewed in each cluster.

[0114] The dividing module 322 can also be configured to re-divide each of the data to be reviewed into each cluster based on the updated centroid vectors until the updated centroid vectors do not change, and determine the data to be reviewed finally divided into each cluster as the data to be reviewed under different categories.

[0115] In a specific application scenario, in order to pre-process the data to be reviewed, the device further includes: a processing unit 36 ​​and an anomaly detection unit 37 .

[0116] The processing unit 36 ​​can be used to determine the same type of data in the data to be reviewed based on the identifiers of each data to be reviewed, and to determine redundant data in the data to be reviewed based on the timestamp information of each data in the same type of data, and to eliminate the redundant data to obtain the processed data to be reviewed.

[0117] The anomaly detection unit 37 can be used to determine the data density within a preset neighborhood corresponding to each data in the processed data to be reviewed, and based on the data density, determine the core data and non-core data in each of the data; take any core data in the core data as a target core data, cluster the remaining core data with the target core data as the cluster center, and obtain the core data under different clustering categories; take any non-core data in the non-core data as a target non-core data, determine the reference core data closest to the non-core data in the core data under each clustering category, and judge whether the distance between the non-core data and the reference core data is less than a preset distance threshold; if the distance between the non-core data and the reference core data is less than the preset distance threshold, the non-core data is divided into the cluster category to which the reference core data belongs, otherwise, the non-core data is determined as abnormal data; the abnormal data is eliminated from the processed data to be reviewed to obtain the cleaned data to be reviewed.

[0118] The classification unit 32 may be specifically configured to classify the cleaned data to be reviewed of the bidder based on each of the review indicators to obtain data to be reviewed in different categories.

[0119] In a specific application scenario, in order to determine the terminal score under each review indicator, the second scoring unit 34 includes a second determination module 341 , a standardization module 342 , a generation module 343 , and a scoring module 344 .

[0120] The second determination module 341 may be used to, when multiple bidders are involved, use any evaluation indicator in each of the evaluation indicators as a target evaluation indicator and determine multiple comparison items under the target evaluation indicator.

[0121] The second determination module 341 may also be used to determine the data to be compared under each comparison item in the data to be reviewed corresponding to the target review indicator.

[0122] The standardization module 342 can be used to standardize the data to be compared for each bidder under each comparison item, so as to obtain the standardized data to be compared.

[0123] The generation module 343 can be used to generate a differential comparison graph of each bidder under the same comparison item based on the standardized data to be compared of each bidder under each comparison item. Wherein, the horizontal axis of the differential comparison graph represents each bidder, and the vertical axis represents the standardized data to be compared of each bidder under the same comparison item.

[0124] The scoring module 344 can be used to send each differential comparison graph corresponding to each evaluation index to the corresponding evaluation terminal, so that each evaluation terminal can determine the terminal score under each evaluation index based on the corresponding differential comparison graph.

[0125] In a specific application scenario, in order to process the data of each bidder, the device further includes a storage unit 38.

[0126] The storage unit 38 can be used to determine the affiliated enterprises of the enterprise to which the target tender project belongs, and determine multiple network servers corresponding to each affiliated enterprise; obtain various data of multiple cooperative bidders in each network server, encrypt the various data, and store the encrypted various data in the cache.

[0127] The acquisition unit 31 can also be used to obtain the encrypted data of the bidder under the project category in the cache based on the project category carried in the tender signal, and obtain the data submitted by the bidder during the current tender process, and the encrypted data and the submitted data constitute the data to be evaluated.

[0128] In a specific application scenario, in order to determine the evaluation result of the bidder, the determination unit 35 includes a third determination module 351 and an addition module 352.

[0129] The third determination module 351 can be used to respectively determine the index weight corresponding to each evaluation index, and determine the scoring weights corresponding to the predicted score and the terminal score respectively.

[0130] The addition module 352 can be used to add the predicted score and the terminal score under each evaluation index based on the scoring weight to obtain the index score under each evaluation index.

[0131] The adding module 352 can also be used to add up the index scores under each review index based on the index weights to obtain the comprehensive score of the tenderer, and determine the review result of the tenderer based on the comprehensive score.

[0132] In a specific application scenario, in order to train and construct a preset scoring model, the device further includes a construction unit 39.

[0133] The construction unit 39 can be used to construct multiple initial scoring models and obtain a sample data set, where the sample data set includes sample review data under different indexes with label information, and the label information is the actual score corresponding to different indexes; divide the sample data set into multiple groups of training data and multiple groups of test data based on the number of models of the initial scoring models; use each group of training data to train the corresponding initial scoring model, and use each group of test data to test the corresponding trained initial scoring model, and determine the initial scoring model that meets the test conditions as the preset scoring model.

[0134] It should be noted that for other corresponding descriptions of each functional module involved in the tender review data processing device provided in the embodiments of the present invention, reference can be made to Figure 1 the corresponding description of the method shown, which will not be repeated here.

[0135] Based on the above as Figure 1 shown in the method, correspondingly, the embodiments of the present invention also provide a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the following steps are implemented: in response to a tender signal of a target tender project, obtain the data to be reviewed of the tenderer, and determine multiple review indexes of the target tender project; classify the data to be reviewed based on each review index to obtain the data to be reviewed under different categories; input the data to be reviewed under each category into a preset scoring model for scoring to obtain the predicted score under each review index; determine the review terminal corresponding to each review index, and send the data to be reviewed corresponding to each review index to the corresponding review terminal respectively, so as to determine the terminal score under each review index through each review terminal based on the corresponding data to be reviewed; receive the terminal score of each review index, and determine the review result of the tenderer based on the predicted score and the terminal score under each review index.

[0136] Based on the above as Figure 1 shown in the method and as Figure 4 shown in the embodiment of the device, the embodiments of the present invention also provide an entity structure diagram of a computer device, as Figure 5As shown in the figure, the computer device includes: a processor 41, a memory 42, and a computer program stored on the memory 42 and executable on the processor. Both the memory 42 and the processor 41 are provided on a bus 43. When the processor 41 executes the program, the following steps are implemented: in response to a tender signal of a target tender project, obtaining data to be reviewed of a tenderer, and determining multiple review indicators of the target tender project; classifying the data to be reviewed based on each of the review indicators to obtain data to be reviewed under different categories; inputting the data to be reviewed under each category into a preset scoring model for scoring respectively to obtain a predicted score under each review indicator; determining a review terminal corresponding to each review indicator, and respectively sending the data to be reviewed corresponding to each review indicator to the corresponding review terminal, so as to determine a terminal score under each review indicator through each review terminal based on the corresponding data to be reviewed; receiving the terminal score of each review indicator, and determining a review result of the tenderer based on the predicted score and the terminal score under each review indicator.

[0137] Through the technical solution of the present invention, the present invention classifies the data to be reviewed of the tenderer through review indicators, and first inputs the data to be reviewed under each category into a preset scoring model for scoring different review indicators. At the same time, the data to be reviewed of different categories are respectively sent to the corresponding review terminals for scoring different review indicators. Finally, according to the scores predicted by the model and the scores given by the review terminals, the review result of the tenderer is determined. Thus, by pre-classifying the data to be reviewed according to the review indicators, when predicting the scores under different review indicators, it is possible to avoid the time and computing resources wasted by analyzing the data to be reviewed that are not relevant to the review indicators, thereby improving the efficiency of score determination, saving computing resources, and further improving the efficiency of review result determination. At the same time, the present invention comprehensively analyzes the scores determined by the review terminals and the model respectively to determine the final review result, which can avoid the limitations of a single evaluation method and the problems of low processing efficiency and low processing accuracy caused by solely relying on manual processing of the review process. Therefore, the present invention can improve the accuracy and objectivity of the tender review process, and further improve the accuracy of the review result.

[0138] Obviously, those skilled in the art should understand that the various modules or steps of the present invention described above can be implemented by a general-purpose computing device. They can be concentrated on a single computing device or distributed on a network composed of multiple computing devices. Optionally, they can be implemented by program codes executable by the computing device. Thus, they can be stored in a storage device and executed by the computing device. And in some cases, the steps shown or described can be executed in a sequence different from that here, or they can be separately fabricated into individual integrated circuit modules, or multiple modules or steps among them can be fabricated into a single integrated circuit module for implementation. In this way, the present invention is not limited to any specific combination of hardware and software.

[0139] The foregoing are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for processing bidding review data, characterized in that: include: In response to a bidding signal of a target bidding project, obtaining data to be evaluated of the bidder, and determining a plurality of evaluation indicators of the target bidding project; Based on each of the review indicators, the data to be reviewed are classified to obtain data to be reviewed in different categories; Input the data to be reviewed under each of the categories into the preset scoring model for scoring, and obtain the predicted score under each of the review indicators; Determine the review terminal corresponding to each of the review indicators, and send the to-be-reviewed data corresponding to each of the review indicators to the corresponding review terminals, so as to determine the terminal score under each of the review indicators based on the corresponding to-be-reviewed data through each of the review terminals; The terminal score of each of the evaluation indicators is received, and the evaluation result of the bidder is determined based on the predicted score and the terminal score under each of the evaluation indicators.

2. The method according to claim 1, characterized in that Based on each of the review indicators, the data to be reviewed are classified to obtain data to be reviewed in different categories, including: Determine the centroid vector of the cluster corresponding to each of the review indicators, and determine the review feature vector corresponding to each of the to-be-reviewed data; Based on the review feature vector and the centroid vector, calculate the distance between each of the to-be-reviewed data and the centroid of each cluster, and divide each of the to-be-reviewed data into each of the clusters based on the distance; Based on the review feature vector corresponding to the to-be-reviewed data in each of the clusters, determining an updated centroid vector corresponding to each of the clusters; Based on the updated centroid vector, each of the to-be-reviewed data is re-divided into each of the clusters until the updated centroid vector does not change, and the to-be-reviewed data finally divided into each of the clusters is determined as to-be-reviewed data under different categories.

3. The method according to claim 1, characterized in that Before classifying the data to be reviewed based on each of the review indicators to obtain the data to be reviewed in different categories, the method further includes: Based on the identifiers of the data to be reviewed, determine the same data in the data to be reviewed, and based on the timestamp information of each data in the same data, determine the redundant data in the same data, and remove the redundant data to obtain the processed data to be reviewed; Determine the data density in a preset neighborhood corresponding to each data in the processed data to be reviewed, and determine the core data and non-core data in each data based on the data density; Taking any core data in the core data as target core data, clustering the remaining core data with the target core data as the cluster center to obtain core data under different clustering categories; Taking any non-core data in the non-core data as target non-core data, determining the reference core data closest to the non-core data in the core data under each clustering category, and judging whether the distance between the non-core data and the reference core data is less than a preset distance threshold; If the distance between the non-core data and the reference core data is less than the preset distance threshold, the non-core data is classified into the cluster category to which the reference core data belongs; otherwise, the non-core data is determined as abnormal data; Eliminating the abnormal data from the processed data to be reviewed to obtain the cleaned data to be reviewed; Based on each of the review indicators, the data to be reviewed are classified to obtain data to be reviewed in different categories, including: Based on each of the evaluation indicators, the cleaned data to be evaluated of the bidder is classified to obtain data to be evaluated in different categories.

4. The method according to claim 1, characterized in that: Each of the review terminals determines a terminal score under each of the review indicators based on the corresponding data to be reviewed, including: In the case of multiple bidders, any evaluation indicator in each of the evaluation indicators is used as a target evaluation indicator, and multiple comparison items under the target evaluation indicator are determined; Determine the data to be compared under each comparison item in the data to be reviewed corresponding to the target review indicator; Standardizing the data to be compared of each bidder under each comparison item to obtain standardized data to be compared; Based on the standardized data to be compared of each bidder under each comparison item, a difference comparison chart of each bidder under the same comparison item is generated, wherein the horizontal axis of the difference comparison chart represents each bidder, and the vertical axis represents the standardized data to be compared of each bidder under the same comparison item; Each of the difference comparison graphs corresponding to each of the review indicators is sent to a corresponding review terminal, so that each of the review terminals can determine a terminal score under each of the review indicators based on the corresponding difference comparison graphs.

5. The method according to claim 1, characterized in that Before obtaining the bidder's data to be reviewed, the method further includes: Determine the subsidiaries of the enterprise to which the target bidding project belongs, and determine a plurality of network servers corresponding to each of the subsidiaries; Obtaining various types of data of multiple cooperative bidders in each of the network servers, encrypting the various types of data, and storing the encrypted various types of data in a cache; The obtaining of the bidder's data to be reviewed includes: Based on the project category carried in the bidding signal, the encrypted data of the bidder under the project category is obtained in the cache, and the data submitted by the bidder during this bidding process is obtained, and the encrypted data and the submitted data constitute the data to be evaluated.

6. The method according to claim 1, characterized in that The step of determining the evaluation result of the bidder based on the predicted score and the terminal score under each evaluation indicator includes: Determine the indicator weight corresponding to each of the review indicators, and determine the score weights corresponding to the predicted score and the terminal score respectively; Based on the scoring weight, the predicted score and the terminal score under each review indicator are added together to obtain an indicator score under each review indicator; Based on the indicator weights, the indicator scores under each of the evaluation indicators are added together to obtain the comprehensive score of the bidder, and based on the comprehensive score, the evaluation result of the bidder is determined.

7. The method according to claim 1, characterized in that Before inputting the to-be-reviewed data under each of the categories into a preset scoring model for scoring and obtaining a predicted score under each of the review indicators, the method further includes: Constructing multiple initial scoring models and obtaining a sample data set, wherein the sample data set includes sample review data under different indicators with label information, and the label information is the actual score corresponding to the different indicators; Based on the model quantity of the initial scoring model, dividing the sample data set into multiple groups of training data and multiple groups of test data; The corresponding initial scoring model is trained using each of the training data, and the corresponding trained initial scoring model is tested using each of the test data, and the initial scoring model that meets the test conditions is determined as the preset scoring model.

8. A device for processing bidding review data, characterized in that: include: An acquisition unit, configured to acquire the to-be-evaluated data of the bidder in response to the bidding signal of the target bidding project, and to determine a plurality of evaluation indicators of the target bidding project; A classification unit, used for classifying the data to be reviewed based on each of the review indicators to obtain the data to be reviewed in different categories; The first scoring unit is used to input the to-be-assessed data under each of the categories into a preset scoring model for scoring, and obtain a predicted score under each of the review indicators; A second scoring unit is used to determine the review terminal corresponding to each of the review indicators, and send the to-be-reviewed data corresponding to each of the review indicators to the corresponding review terminals, so as to determine the terminal score under each of the review indicators based on the corresponding to-be-reviewed data through each of the review terminals; A determination unit is used to receive the terminal score of each of the evaluation indicators, and determine the evaluation result of the bidder based on the predicted score and the terminal score under each of the evaluation indicators.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images